Positioning scanning is usually needed when a leg is scanned to observe a specific part. Many positioning scanning methods are currently available. There is a technology based on three-dimensional (3D) data in which a 3D image of a leg in a predetermined positioned scanning direction is obtained by a 3D imaging device, an approximate position of a knee joint in the generated image is determined, a volume of interest at the approximate position is defined, and the volume of interest divided to obtain the image of the thighbone on the knee joint, a posterior condyle line (PCL) is detected, and a positioning scanning is performed on the leg.
However, the technology based on 3D data spends a long time to obtain 3D data in most cases, and the segmentation based on 3D data is extremely sensitive to noises, segmentation thresholds, image quality and scanning parameters and the like.
Therefore, a technology capable of addressing the above problem is desired.